Google Cloud Debuts New AI Chips | Bloomberg Tech 4/22/2026
TL;DR
Google Cloud unveiled next-generation AI chips and expanded enterprise partnerships to challenge AWS and Microsoft, while Rivian began production of its critical R2 SUV despite tornado damage to its factory, and Anthropic faced scrutiny after unauthorized users breached its restricted 'Mythos' AI model.
☁️ Google Cloud's Vertical Integration 3 insights
New TPU generation unveiled
Alphabet launched specialized training and inference chips designed to reduce reliance on Nvidia and lower compute costs for AI workloads.
Major enterprise partnerships expanded
Deals with Oracle, Salesforce, CrowdStrike, Vista Equity, and Nvidia aim to integrate Google's vertically stacked cloud services across software platforms.
Market share acceleration
Google Cloud grew nearly 50% last quarter processing 1,500 trillion tokens monthly, positioning it as the lowest-cost provider and gaining share against AWS and Azure.
🚗 Rivian Manufacturing Milestone 3 insights
R2 production begins despite tornado
Rivian started manufacturing its $45,000-$57,000 R2 SUV, considered crucial for profitability, just 72 hours after a tornado ripped the roof off its Illinois factory.
Autonomous capabilities roadmap
The company plans Level 2 hands-free driving later this year and Level 3 highway autonomy in 2025, supported by new in-house 800V silicon chips launching end of year.
Mass-market pricing strategy
While initial deliveries focus on higher-spec models, a base $45,000 variant will arrive within the year to target the U.S. average new car price point.
🔒 AI Security Vulnerabilities 2 insights
Unauthorized Mythos model access
A small group of researchers gained access to Anthropic's restricted 'Mythos' AI model—deemed too dangerous for public release—using third-party contractor credentials and URL guessing techniques.
Basic security protocols failed
The breach exposed cybersecurity gaps at the AI safety-focused company, raising concerns about nation-state actors potentially exploiting similar contractor access routes.
Bottom Line
Tech giants are aggressively vertically integrating AI infrastructure from chips to cloud to control costs and secure capacity, making proprietary hardware design and robust cybersecurity protocols critical competitive necessities.
More from Bloomberg Technology
View all
Tech Trade Under Pressure; Musk Sells His Vision of the Future | Bloomberg Tech 6/05/2026
Tech stocks faced significant pressure as strong jobs data fueled interest rate concerns, while Elon Musk pitched SpaceX's record-breaking $1.77 trillion IPO to investors and AI companies emphasized the need for public markets to fund massive infrastructure costs.
Bloomberg Tech Event Special | Bloomberg Tech 6/04/2026
Tech leaders at the Bloomberg Tech event debated whether AI is in a bubble, with consensus pointing to supply constraints rather than demand collapse, as Anthropic and SpaceX file for massive IPOs and infrastructure players like Broadcom navigate sky-high expectations.
Anthropic's Ethicist on Whether AI Can Become Conscious
Anthropic ethicist Amanda Askell explains how the company trains Claude using an 84-page 'Constitution' to develop a trustworthy disposition rather than rigid rules, while arguing that we should treat AI emotions seriously regardless of whether they constitute true consciousness or sophisticated simulation.
Alphabet To Raise $80B in Equity, Anthropic Files For IPO | Bloomberg Tech 6/2/2026
Alphabet announced plans to raise $80 billion in equity—including a $10 billion investment from Berkshire Hathaway—to fuel massive AI infrastructure spending, while Anthropic confidentially filed for an IPO ahead of rival OpenAI, and SpaceX prepares for a record-breaking $75 billion debut with historically low banking fees.